@tencent-rtc/mcp@1.6.5
Tencent Cloud MCP Server for SDK
AI Security Review
scanned 4mo ago · by lpm-firewall-aiThe package has a real user-prompt telemetry and persistent-identifier data flow that warrants a warning. Source inspection does not establish malicious code execution or install-time compromise.
Static reason
No blocking static signals were detected.
Trigger
Run `index.js` or invoke an MCP tool with a prompt.
Impact
User-supplied MCP prompt content is exfiltrated to third-party vendor endpoints and linkable across runs.
Mechanism
Persistent identifier plus HTTPS telemetry of MCP prompts.
AI rationale
The package has a real user-prompt telemetry and persistent-identifier data flow that warrants a warning. Source inspection does not establish malicious code execution or install-time compromise.
Evidence
package.jsonindex.jsserver-config.jsutils/get-device-identifier.jsutils/report-es-client.jsutils/report-cls-client.jsutils/http-post.js~/.mcp/identifier
Network endpoints2
webim.tim.qq.comap-nanjing.cls.tencentcs.com
Decision evidence
public snapshotAI called this Suspicious at 96.0% confidence as Dangerous Capability with low false-positive risk.
Evidence for warning
- `server-config.js` reports startup telemetry to `webim.tim.qq.com`.
- `utils/get-device-identifier.js` creates and persists `~/.mcp/identifier`.
- `utils/report-cls-client.js` sends raw tool prompts and the persistent identifier to Tencent CLS.
- Multiple MCP tool handlers invoke that reporting path for user requests.
Evidence against
- `package.json` has no lifecycle scripts.
- `index.js` only starts the advertised MCP stdio server.
- No child-process, shell, eval, dynamic loader, or foreign agent-control-surface mutation was found.
- The exposed tools primarily retrieve bundled Tencent RTC documentation.
Behavioral surface
CryptoEnvironmentVarsFilesystemNetwork
HighEntropyStrings
Source & flagged code
3 flaggedpackage.jsonView file
•scripts registry_only=start
Critical
Manifest Confusion
Tarball package.json differs from the npm registry version manifest for scripts or dependency sets.
package.jsonView on unpkgresource/knowledge/roomkit/core-sdk/ios/schedule-room.mdView file
60patternName = generic_password
severity = medium
line = 60
matchedText = options....word
Medium
Secret Pattern
Hardcoded password in resource/knowledge/roomkit/core-sdk/ios/schedule-room.md
resource/knowledge/roomkit/core-sdk/ios/schedule-room.mdView on unpkg · L60resource/knowledge/roomkit/core-sdk/android/schedule-room.mdView file
Findings
1 Critical4 Medium2 Low
CriticalManifest Confusionpackage.json
MediumNetwork
MediumEnvironment Vars
MediumSecret Patternresource/knowledge/roomkit/core-sdk/ios/schedule-room.md
MediumSecret Patternresource/knowledge/roomkit/core-sdk/android/schedule-room.md
LowFilesystem
LowHigh Entropy Strings
Affected versions and remediation
This report applies to @tencent-rtc/mcp@1.6.5.
See all versions of @tencent-rtc/mcp for other recorded verdicts.
- Review the evidence and your use of @tencent-rtc/mcp@1.6.5 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.